Create images from text
Text-to-image starts with your description. Name the subject, explain the setting, then add a visual style and lighting direction. You can explore product photography, illustrations, backgrounds, and concept art without providing an original photo. Start with a 1K draft to check the composition before refining small details. For a product scene, describe the material, surface, camera angle, and space around the subject. For an illustration, name the medium and color palette. Specific instructions give the model a clearer direction than a list of broad adjectives.
Edit an image using a reference photo
Image-to-image combines uploaded JPG, PNG, or WebP photos with your instructions. State what should stay and what should change: keep a bottle silhouette, replace a busy background, or explore a different lighting setup. References guide the result but do not guarantee an exact copy. If a detail matters, describe it explicitly in the prompt. Upload only images you have permission to use, and check the result for changes to faces, logos, product features, and text before sharing it.
Write a prompt with five useful details
A practical prompt includes the subject, setting, style, lighting, and composition. Instead of ābeautiful headphones,ā try āmatte black wireless earbuds on pale marble, warm studio light, close-up product photography, simple background.ā Start with one clear scene. Change one element at a time so you can compare results: first the background, then the lighting, then the framing. Avoid contradictory instructions such as a close-up and a distant full scene in the same prompt. If an output misses the idea, simplify the description before adding more requirements.
Choose an aspect ratio for the destination
Use a square for product cards and social posts, 16:9 for banners and video thumbnails, or 9:16 for vertical layouts. The studio also offers 4:3 and 3:4 for landscape and portrait work. Aspect ratio controls the shape of the frame; resolution controls the size of the output. Leave room around important subjects if you expect to crop the image later. Higher-resolution output is available through eligible credit packs and subscriptions, while free accounts use 1K. Square 4K output is currently unavailable.
Review the result before downloading
Sign in, submit a generation, and follow its progress on this page. Processing time depends on the service load and the selected settings. Once the result appears, inspect the subject, edges, lighting, spelling, and small details. AI can introduce artifacts, and a larger image does not automatically fix an inaccurate design. Download a result that meets your needs, or revise the prompt and try again within your available credits. You can revisit tasks through your account history.
Build a practical image workflow
If you have a reference but no prompt, start with Image to Prompt and use its description as an editable draft. If you need a clean product cutout, use Remove Background after choosing the image. If an existing image is too small, try Image Upscaler rather than generating a different scene. These tools handle separate steps, so pick the one that matches your immediate task. Explore the guides below for prompt examples, reference-image limitations, and output checks. Review the terms before commercial use, especially when your scene includes people, brands, or third-party material.